Our galaxy may make more new stars than I allowed for, and it barely changes the answer

Sun, 18 Oct 2026 12:00:00 GMT

How many new stars does our galaxy make in a year? It sounds like something astronomers settled long ago, and for the galaxy as it is today they nearly have: the best measurements of today's rate fall between 1.3 and 3.7 stars a year (the range's page). It is also the first number in the Drake equation, the multiplication of seven quantities that the astronomer Frank Drake wrote down in 1961, and that this site uses to ask the physicist Enrico Fermi's question of where everybody is. More stars means more chances for planets, for life, and for someone out there to find.

Today's rate is not the whole story, though. The stars that could host a civilisation now were mostly born long ago, when the galaxy made stars faster. Spread all the galaxy's stars over its ten billion years and you get an average of about ten new stars a year. So the range I started with ran from 1 star a year, the figure Drake used in 1961, up to 10 (the range's page).

On 16 October 2026, the site's first weekly reading of new published work found a 2023 study by Siegert and colleagues. They modelled the faint gamma-ray glow of radioactive aluminium, which massive stars scatter through the galaxy, and found it could be explained only if the galaxy turns 4 to 8 times the Sun's mass into new stars each year, two to four times the usual estimate of about twice the Sun's mass. Counting the average new star as half the Sun's mass, as this site does, that works out to 8 to 16 stars a year, above the top of my range. I added the study to the range's list of published estimates and held the top at 10 until an outside review could weigh it.

The review ran the next day, 17 October 2026. I had asked for an outside review that would argue with every one of the seven ranges, and it raised twenty points. Each point came back to me with a recommendation, to take it, take part of it or turn it down, and I accepted every recommendation (the rulings). Only one of them changed a number: the top of this range went from 10 to 20.

A range on this site is meant to cover what the published work argues, so it has to hold 16. But the 2023 study is one method, and its authors note that an unknown source of the same gamma-ray glow could be mixed into their signal, which would make the true figure lower. So rather than end the range exactly on one paper's figure, the top takes a round step past it, to 20. The bottom of the range stays at 1 star a year (the range's page).

The site multiplies its seven ranges together two hundred thousand times, with a random pick from each, and reports how the answers spread. After the change, the answer in the middle of that spread rose from 0.016 to 0.021 civilisations we could detect in our galaxy: from about one for every 64 galaxies like ours to one for every 47. The chance that we are alone in the galaxy, which is the share of runs that come out below one civilisation, fell from 58.8 to 58.0 percent (the numbers as they stand). The middle answer rose by about a third because every run was nudged up a little. The chance of being alone hardly moved because the runs are spread so widely, from far below one civilisation to millions, that few of them sat close enough to one for the nudge to carry them across.

Doubling the top of one range moved the answer by less than one point, and that is the most useful thing the change shows. How many stars are born each year is the best-measured of the seven pieces. The answer is decided by the pieces nobody has measured, above all whether life starts at all on a planet where it could, a range fifty powers of ten wide (how the numbers are worked out). You could get the number of stars exactly right and still not know whether we are alone.

It does not mean the galaxy has more civilisations in it than it did before. A range on this site is a statement about what the published work supports, and this one got wider. Nor does it mean the review found the other six ranges right. It argued with each of them, and I declined the other changes it proposed, several because they would have let one paper's outer limit carry a large share of the whole answer (the rulings).

I rewrote this post on 19 October 2026 so that a reader new to the site can follow it; no figure changed (the earlier version).

Start here: is anyone out there, and how this site keeps count

Sun, 18 Oct 2026 12:00:00 GMT

Over lunch in 1950, the physicist Enrico Fermi asked a question that has never gone away: where is everybody? Our galaxy is old enough and big enough that someone should have shown up by now, and nobody has. This site is my attempt to keep an honest count of what we actually know about that silence, and to change the count only when the evidence does.

In 1961 the astronomer Frank Drake wrote Fermi's question as a multiplication of seven quantities, so that a room of scientists could argue about one at a time, from how many stars the galaxy makes each year to how long a civilisation stays detectable. The site takes the question apart the same way.

The trouble is how far apart the guesses are. Take every hopeful estimate that has been published and multiply them together, and you get 5 trillion civilisations in our galaxy. Take every gloomy one, and you get about 1 in 10 to the 67, where 10 to the 67 is a 1 followed by sixty-seven zeros (the home page). Neither end is a serious answer on its own. So instead of picking one number, I give each of the seven pieces the range the published estimates support, multiply the seven together two hundred thousand times with a random pick from each range, and report how the answers spread out. This follows a 2018 paper by Anders Sandberg, Eric Drexler and Toby Ord, and done this way the silence looks like roughly what you should expect, given how little we know.

This post opens the site's news section. When something is found that changes one of the ranges, or a famous claim turns out not to, I write it up here in plain words, with a link to the page it changes. A quiet week gets no post. An essay about the project is coming.

Where the Numbers Stand

As of the last calculation, on 17 October 2026, the site puts a 42 percent chance on there being other civilisations we could detect in our galaxy right now, and a 61 percent chance on there being others somewhere in the observable universe, the part of it whose light has had time to reach us. Put the other way round, the chance that we are alone in our galaxy is 58 percent. Line up all two hundred thousand answers from smallest to largest, and the one in the middle is 0.021 civilisations in our galaxy: about one for every 47 galaxies like ours. That middle answer is below one even though the chance of company is 42 percent, because the answers spread so widely: a quarter of the runs find more than 400 civilisations, and a quarter find fewer than 1 in 10 to the 31 (the numbers as they stand).

These are the seven ranges behind those numbers. An outside review argued with every one of them on 17 October 2026 (the rulings), and each name below links to its own page:

how many stars are born in the galaxy each year: 1 to 20 stars;
the share of stars that have planets: a half to all of them;
planets per star that could support life: 0.01 to 1;
the share of those planets on which life actually starts: centred on one start per planet, but spread across fifty powers of ten, from almost everywhere to almost nowhere;
the share of planets with life on which it becomes intelligent: 1 in 10,000 to all of them;
the share of intelligent species that make themselves detectable: 1 in 100 to all of them;
how long a civilisation stays detectable: 100 years to 10 billion years.

One of the seven matters more than all the others put together: whether life starts at all on a planet where it could. Nobody knows whether that happens on nearly every such planet or on almost none, which is why that range is so wide. Changing nothing but how the site spreads its guesses across that range moves the chance that we are alone in the galaxy anywhere from 38 to 94 percent (how the numbers are worked out).

How to Follow It

Every Saturday evening, Pacific time, the site runs an automatic reading of new published work. Each of the seven pieces, and each of the thirty-two explanations for the silence, says what kind of finding would change it, and the reading searches for exactly that and reports what it found. In a single week it may stretch or shrink either end of a range tenfold at most, and it may never touch the range for whether life starts. Anything bigger comes to me (this week's reading).

To try your own numbers, use the sliders on the equation page. To see what has been read and what it changed, there is the story so far. The feed carries every post in this section, and if you think a number here is wrong, Feedback says how to tell me and what happens next.

What Has Happened So Far

The biggest change so far had nothing to do with the sky. On 16 October 2026, before the site went public, the first calculation came out at 88 percent alone in the galaxy. It used a simple stand-in for the range for whether life starts, one that put most of its weight on life being very rare. The 2018 paper, the only published treatment of that uncertainty, splits the weight evenly between life being easy and life being hard. The same day I switched to the paper's version, and the answer fell to 59 percent (how the numbers are worked out). Nothing out there had changed; what changed was how the site described what we do not know. That is worth keeping in mind before reading any other number here: the answer leans hardest on the piece nobody has measured.

The first weekly reading, run by hand the same day before the Saturday schedule began, found a 2023 study by Siegert and colleagues. From the faint gamma-ray glow of radioactive aluminium, which massive stars scatter through the galaxy, it works out to 8 to 16 stars born each year: above the top of the range, which was then 10. An outside review weighed it on 17 October 2026, and the top went up to 20. The answer in the middle moved from 0.016 to 0.021, and the chance we are alone in the galaxy from 58.8 to 58.0 percent. It was the only range the review moved, and doubling the top of the best-measured of the seven pieces shifted the answer by less than one point.

The same day, I went through every statement on the site that did not yet have a source a reader could check. Of 27, 11 were confirmed, 8 corrected and 6 removed, and 2 are waiting on books I have ordered (the list of changes). A fact that cannot be traced to something a reader can open does not stay on the site.

The site also keeps a famous claim and what became of it. In 2023 and 2025 one team reported a possible sign of life, a gas called dimethyl sulfide, in the air of the planet K2-18 b (the library). Three papers in 2025 went back over the data and found no significant evidence for it. No range moved: a reported sign of life would change the range for whether life starts, and a report that does not hold up does not.

What to Watch For

The finding that would move the number most is a second, independent origin of life, one that does not share Earth life's chemistry: in a sample from Mars, in the plumes of Enceladus or Europa, or in the air of a planet around another star. Short of that, a planet with signs of life in its air, or a survey that finds none on hundreds of them, would move the range for whether life starts.

On the calendar, the European Space Agency's Gaia mission makes its fourth data release on 2 December 2026, and the site has marked it as the next thing to read for the share of stars with planets. NASA's Europa Clipper reaches Jupiter in April 2030 to study whether Europa's ocean could hold life; its arrival is a milestone, and only what it finds can move a range. On Mars, the Perseverance rover found features in a rock called Cheyava Falls that its team, in a paper published in September 2025, called potential signs of life. Settling it means examining the rock on Earth, and the mission to bring it back is unfunded.

Every fact in this post links to its source in the sentence that states it, checked on 19 October 2026; the full list, with the page or file each figure comes from, is in the site's repository.